Billions of dollars into the robotics industry, there is still no agreed-upon, buyer-facing way to measure what a robot you can actually purchase will actually do. Lab benchmarks are built for researchers. Manufacturer spec sheets are built for marketing. Procurement happens in the gap.
BotInfo's evaluation standard scores buyable robots on the things buyers ask us about: how long until it runs, how long it runs for, what it can carry, what it legally is in the United States, what it costs to get to operational — with every score carrying an evidence grade and a date.
No score on this site is ever stronger than its evidence. Every published figure carries one of four grades:
| Grade | Meaning |
|---|---|
| E0 | Spec-sheet claim — manufacturer documentation, cross-checked against regulatory filings where possible. A claim, clearly labeled as one. |
| E1 | Video-verified — scored from dated public footage, source cited. |
| E2 | Hands-on — a BotInfo-operated unit run through our published test protocol. |
| E3 | Field report — verified feedback from an institutional deployment. |
Grades are never blurred. An E1 score is “video-verified,” never “tested by us.” Where the evidence doesn't support a score, the scorecard says “insufficient evidence” — an honest gap beats an invented number, and the gaps themselves tell you what no one in this industry will currently show you.
Each robot is assessed against nine buyer-relevant categories:
Categories 6 and 9 draw on datasets BotInfo already maintains daily: verified checkout pricing with visible dates, and per-model US authorization status — data no vendor publishes honestly about itself.
Regulatory scoring rubric (category 6): 5 = per-model FCC grant verified in the public record · 4 = authorized-generation platform (grantee-level records, model mapping not 1:1 public) corroborated by flowing US distribution · 2 = authorization status unresolved · 0 = not authorized for US import/marketing/sale.
Each scorecard below is scored against the battery above, with every line graded and dated.
Published scores are never for sale and never vendor-influenced. Vendors may in future pay for evaluation services — expedited hands-on testing — but never for outcomes. If that line ever moves, this page will say so in plain text, dated.
.png)
